The pattern, and the thing hiding underneath it

Both charts have the same thing going up the side – how much a university's teaching score improved between the 2023 and 2026 surveys. On the left, that improvement is set against the change in staffing. On the right, against how low the university started in 2023. The right-hand line is far steeper, and that is the problem.

1. The apparent staffing effect

Change in the ratio each cohort actually studied under, against the gain in teaching score

2. The pandemic rebound

Where the university stood in 2023, against the same gain in teaching score

Universities that scored badly in 2023 improved most, because they had the furthest to climb back after the pandemic. That rebound is close to three times stronger than anything staffing does, and it is mixed into the left-hand chart. Any claim about staffing has to be made after taking it out.
Teaching score gain 2023 to 2026, by how much the experienced ratio changed
GroupChange in ratioTeachingLearning opportunitiesAssessmentAcademic supportOrganisationResourcesStudent voice

Each dot is one university, 106 in all, with at least 100 NSS responses in both years. Teaching score is NSS Theme 1. The experienced ratio is the average staff to student ratio across the years each cohort studied, from the Guardian University Guide. Correlation is not cause – a worsening ratio usually means student numbers grew faster than staff, which happens at universities under financial pressure that are doing many other things at once.